14:01:03 <jlibosva> #startmeeting networking
14:01:04 <openstack> Meeting started Tue Mar 13 14:01:03 2018 UTC and is due to finish in 60 minutes.  The chair is jlibosva. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:01:07 <jlibosva> hi all
14:01:08 <openstack> The meeting name has been set to 'networking'
14:01:11 <mlavalle> o/
14:01:13 <jlibosva> #topic Announcements
14:01:17 <aspiers> o/
14:01:34 <jlibosva> This is the first Tuesday meeting after the PTG
14:01:39 <pasuder> hello!
14:01:41 <slaweq> hello
14:01:43 <jlibosva> mlavalle sent out a nice summary here
14:01:45 <jlibosva> #link http://lists.openstack.org/pipermail/openstack-dev/2018-March/128183.html
14:01:49 <pasuder> true +1
14:02:06 <haleyb> hi
14:02:13 <ndefigueiredo> Hi guys
14:02:29 <jlibosva> in development cycle, we're currently at Rocky-1 which is due at the week of 16 Apr
14:02:33 * mlavalle wonders how haleyb is doing wit the latest storm
14:02:49 <hichihara> hi
14:02:57 <haleyb> no trees fell down yet, cross fingers
14:03:10 <mlavalle> good luck!
14:03:23 <jlibosva> fingers crossed haleyb :)
14:03:35 <amotoki> hi
14:03:42 <jlibosva> that's all from me, any other announcements from anybody?
14:03:46 <rubasov> hi
14:04:01 <mlavalle> Not from me
14:04:26 <ndefigueiredo> just a question about the meeting structure
14:04:38 <ndefigueiredo> is there an agenda?
14:04:51 <jlibosva> ndefigueiredo: we have a wiki page: https://wiki.openstack.org/wiki/Network/Meetings
14:05:04 <ndefigueiredo> thanks
14:05:20 <mlavalle> ndefigueiredo: good question :-)
14:05:53 <mlavalle> ndefigueiredo: and you can add a topic at the end, in the open agenda section
14:06:02 <mlavalle> if you need to
14:06:03 <jlibosva> and I need to remove the one that I added :)
14:06:10 <jlibosva> as it was discussed already
14:06:13 <jlibosva> anyway, moving on
14:06:17 <jlibosva> #topic Blueprints
14:06:18 <mlavalle> you are always welcome to add topics
14:06:28 <jlibosva> #link https://launchpad.net/neutron/+milestone/rocky-1
14:07:11 <mlavalle> Regarding the Rocky-1 milestone, I want to point out that it is not complete yet
14:07:11 <ndefigueiredo> mlavalle: will do.
14:07:21 <mlavalle> in the dashboard
14:07:44 <jlibosva> ack
14:08:17 <mlavalle> I have to go thorugh the PTG summary
14:08:41 <mlavalle> and add blueprints from there that don't exist in Launchpad yet
14:08:45 <pasuder> mlavalle: I noticed in the morning that I am assigned person to subnetpool blueprint. Shall I do something more or just continue with working on implementation?
14:09:04 <mlavalle> pasuder: just continue working on implementation
14:09:17 <pasuder> mlavalle: thx
14:09:22 <mlavalle> I am trying to do the "paperwork" for you all
14:09:32 <amotoki> pasuder: which one?
14:09:35 <mlavalle> so you can focus on the development side
14:09:39 <pasuder> thx!
14:09:46 <pasuder> amotoki: https://blueprints.launchpad.net/neutron/+spec/subnet-onboard
14:09:46 <mlavalle> the Launchpad "paperwork"
14:09:57 <pasuder> mlavalle: understand, thank you
14:10:12 <amotoki> pasuder: thanks
14:10:44 <mlavalle> as I go through the PTG summary, I might contact some of you when creating the blueprints in Launchpad
14:11:17 <mlavalle> pasuder: the idea of this ection of the meeting is that you can report status of your assigned blueprint
14:11:31 <mlavalle> or ask for hel (reviews for example) in case you are stuck
14:11:37 <mlavalle> help^^^
14:12:01 <pasuder> mlavalle: at this moment I do not have anything to report
14:12:33 <mlavalle> ndefigueiredo: please look at the PTG summary. there is one or two places where your name pops up
14:12:57 <mlavalle> so you and I will have to create one or two blueprints
14:13:08 <mlavalle> abd ad it to the dashboard
14:13:24 <ndefigueiredo> mlavalle: just having some trouble finding things as I learn.
14:13:38 <mlavalle> no problem, we are here to help
14:13:49 <jlibosva> ndefigueiredo: in any case you have troubles, feel free to ping us on #openstack-neutron channel :)
14:13:52 <mlavalle> don't by shy about asking questions, please
14:14:08 <jlibosva> ++
14:14:08 <slaweq> ++
14:14:22 <amotoki> ++
14:14:26 <mlavalle> ndefigueiredo: yeah, jlibosva is in Prague, very close to your time zone
14:14:44 <ndefigueiredo> ok, thank you. If I can ask now: 1) is there a place for adding the agenda items? 2) where can I see the PTG summary?
14:14:48 <mlavalle> the same as slaweq
14:14:54 <mlavalle> he is in Poland
14:15:12 <jlibosva> ndefigueiredo: sure, at this wiki, once you're logged in, you can click Edit button: https://wiki.openstack.org/wiki/Network/Meetings
14:15:31 <jlibosva> ndefigueiredo: and at the very bottom, there is On Demand Agenda where you can put your topic
14:15:40 <ndefigueiredo> jlibosva: cool, thanks :)
14:15:46 <jlibosva> we usually have 5-15 minutes at the end of each meeting dedicated for such topics
14:16:12 <ndefigueiredo> fwiw I am in South Africa (+2 timezone) currently.
14:16:28 <pasuder> ndefigueiredo: same in PL and CZ :)
14:16:41 <mlavalle> and there are no stupid questions.... well, ocasionally that PTG guys comes up with a few
14:16:52 <mlavalle> ;-)
14:17:16 <pasuder> :D
14:17:17 <amotoki> you are in a same time zone even though you are in different seasons :)
14:17:53 <pasuder> amotoki: true :) same but different
14:18:17 <jlibosva> so unless there is anything else to be added to blueprints, I suggest to move on :)
14:18:37 <ndefigueiredo> amotoki: Yes! It is hot here (end of summer)!
14:18:37 <jlibosva> ndefigueiredo: oh, one more thing :)
14:19:02 <jlibosva> ndefigueiredo: even if at the end at the open agenda topic there is nothing in the wiki, feel free to bring any topic even without writing it to the wiki page
14:19:21 <jlibosva> ok, now moving on :)
14:19:23 <jlibosva> #topic Community Goals
14:19:37 <jlibosva> #link https://governance.openstack.org/tc/goals/index.html#release-cycles
14:20:32 <jlibosva> so seems like we have some debts from previous cycles
14:20:46 <amotoki> we newly added this section in Rocky so that we can sync more with community goals.
14:20:59 <mlavalle> here, I want to say that annp sent me an email asking if he can work on https://governance.openstack.org/tc/goals/pike/deploy-api-in-wsgi.html.
14:21:13 <amotoki> :)
14:21:15 <jlibosva> right and I saw at the BP dashboard we already track some of the missing goals
14:21:19 <ndefigueiredo> jlibosva: Cool, got it
14:21:31 <jlibosva> annp++ great :)
14:21:34 <mlavalle> amotoki: you ok with it?
14:21:57 <amotoki> mlavalle: yes
14:22:11 <mlavalle> I'll respond to his email copying you
14:22:15 <amotoki> IIRC we already have a WIP devstack patch. we can resume the work based on it
14:22:35 <mlavalle> yes, please respond to the email with a pointer to it
14:22:46 <amotoki> sure
14:23:33 <jlibosva> I'm looking that the python3.5 support is not written at the wiki page but I don't think we're done with it
14:23:47 <mlavalle> oh really?
14:23:59 <jlibosva> we did some work to make functional tests working with python3 and I think it hasn't been finished
14:24:04 <slaweq> I only know we had some experimental jobs for py35
14:24:18 <jlibosva> https://etherpad.openstack.org/p/py3-neutron-pike
14:24:34 <mlavalle> ok, let's add it to the wiki
14:24:44 <amotoki> so, do we need to enable python3 in more gate jobs?
14:24:55 <mlavalle> and I will hunt down the blueprint if one ever existed and add it to the dasjboard
14:25:14 <jlibosva> I think it would be nice to get at least functional python3 job running and passing
14:25:31 <jlibosva> I'll search for the eventlet bug to see how it stands, I remember vstinner working on it
14:26:01 <mlavalle> jlibosva: was there ever a launchpad blueprint for it? do you know?
14:26:22 <jlibosva> mlavalle: I think there was an Issue on eventlet github IIRc
14:26:39 <mlavalle> ok, I'll also try to find it
14:27:17 <jlibosva> mlavalle: oh, there was: https://bugs.launchpad.net/oslo.service/+bug/1705047
14:27:18 <openstack> Launchpad bug 1705047 in oslo.service "Race condition in signal handling on Python 3" [High,In progress] - Assigned to Victor Stinner (victor-stinner)
14:27:41 <mlavalle> thanks!
14:28:00 <mlavalle> jlibosva: good catch about us missing that community goal as well
14:28:05 <jlibosva> right, I think we're not able to gracefully shutdown multiprocessed processes, like neutron-server
14:28:31 <jlibosva> mlavalle: thanks :)
14:29:00 <amotoki> some updates on the other community goals
14:29:21 <amotoki> re: policy-in-code, there is no explicit progress this week but I had time to look into more. i will update it later this week
14:29:39 <amotoki> re: mox removal, we successfully got assignees :)
14:29:48 <mlavalle> yaay!
14:30:02 <mlavalle> who are they?
14:30:22 <haleyb> this bug is tracking the work, https://bugs.launchpad.net/bugs/1753504
14:30:23 <openstack> Launchpad bug 1753504 in python-neutronclient "Remove mox/mox3 usage from testing" [High,In progress] - Assigned to Hongbin Lu (hongbin.lu)
14:30:23 <amotoki> hongbin works on neutronclient
14:30:30 <amotoki> haleyb: yes
14:30:32 <mlavalle> ah great!
14:30:50 <amotoki> I covered dashboard stuffs too
14:31:13 <amotoki> midonet stuff depends on neutronclient and yamamoto is aware of it too
14:32:31 <slaweq> I can take a look at https://governance.openstack.org/tc/goals/rocky/enable-mutable-configuration.html if there is no other volunteer
14:32:32 <amotoki> that's all from me
14:32:58 <mlavalle> slaweq: thanks. yes, please go ahead
14:33:00 <amotoki> slaweq: thanks.
14:33:19 <jlibosva> thanks for updates
14:33:25 <jlibosva> anything else to community goals?
14:33:29 <mlavalle> so at this point we have asignees for all of them, right amotoki ?
14:33:30 <amotoki> so, we seems to have assignees for all goals
14:33:35 <slaweq> ok, so I will take care, thx
14:33:41 <pasuder> :D
14:33:43 <mlavalle> Great!
14:33:49 <amotoki> let's update the wiki page :)
14:34:21 <jlibosva> good :)
14:34:31 <jlibosva> moving on
14:34:33 <jlibosva> #topic Bugs
14:34:44 <jlibosva> haleyb sent an email about last week
14:34:50 <haleyb> http://lists.openstack.org/pipermail/openstack-dev/2018-March/128182.html
14:34:58 <jlibosva> #link http://lists.openstack.org/pipermail/openstack-dev/2018-March/128182.html
14:35:00 <jlibosva> ah
14:35:02 <mlavalle> you were preparing for the storm, right?
14:35:04 <jlibosva> thanks haleyb :)
14:35:10 * haleyb had that paste waiting :)
14:35:30 <haleyb> i'll go through high and those that need attention
14:35:40 <haleyb> https://bugs.launchpad.net/bugs/1754062
14:35:41 <openstack> Launchpad bug 1754062 in OpenStack SDK "openstack client does not pass prefixlen when creating subnet" [Undecided,New]
14:36:09 <haleyb> two fixes proposed for that, one has merged for the api-ref, openstacksdk change yet to merge
14:36:35 <haleyb> https://bugs.launchpad.net/bugs/1754327
14:36:36 <openstack> Launchpad bug 1754327 in neutron "Tempest scenario jobs failing due to no FIP connectivity" [High,Fix released] - Assigned to Slawek Kaplonski (slaweq)
14:36:40 <amotoki> do we need OSC fix in addition to SDK?
14:37:01 <amotoki> i haven't checked it yet.
14:37:20 <haleyb> amotoki: no, just the sdk it seems, it was filtering the argument from what i remember
14:38:02 <amotoki> haleyb: ah, it is possible. i will check it later.
14:38:31 <haleyb> amotoki: thanks, i knew you had reviewed that change too
14:39:21 <haleyb> moving on the tempest bug, slawek's change has mergeed to increase ssh timeout value, so it is fixed
14:39:46 <haleyb> https://bugs.launchpad.net/bugs/1755243
14:39:47 <openstack> Launchpad bug 1755243 in neutron "AttributeError when updating DvrEdgeRouter objects running on network nodes" [High,In progress] - Assigned to Daniel Gonzalez Nothnagel (dgonzalez)
14:39:52 <slaweq> yes, and I think that lb scenario job is much better now :)
14:40:13 <haleyb> slaweq: yes, thanks!
14:40:16 * jlibosva refreshes grafana
14:40:35 * mlavalle suspected jlibosva was going to do that
14:40:42 <slaweq> LOL
14:40:45 <mlavalle> :-)
14:40:46 <jlibosva> :)
14:41:08 <amotoki> someone seems to have disabled auto refresh :p
14:41:54 <haleyb> regarding the DVR bug ^^^, there was a scenario where the agent could try and read an invalid attribute, fix proposed yesterday, just needs updates based on comments
14:42:00 <haleyb> https://review.openstack.org/#/c/552097/
14:42:08 <jlibosva> wow! 4% at LB job, *clap clap* slaweq
14:42:21 <slaweq> thx
14:42:24 <mlavalle> Nice!
14:42:38 * haleyb waits for applause to lessen before moving on :)
14:43:21 <amotoki> #thanks slaweq succeeds to improve linuxbridge failure rate much :)
14:43:44 <amotoki> no thanks bot?
14:44:03 <haleyb> ok, last high priority from last week is https://bugs.launchpad.net/bugs/1754563
14:44:05 <openstack> Launchpad bug 1754563 in neutron "Arp_responder function has failed since Ocata" [High,Confirmed]
14:45:00 <haleyb> there is a missing setup_privsep() call in the linuxbridge agent, which slaweq had actually just fixed in another change, i cherry-picked the relevant file to queens and will propose to pike as well
14:45:34 <slaweq> haleyb: I think it is also needed for ocata
14:45:47 <slaweq> as there are privileged.add_neigh_entry() used there also
14:46:05 <slaweq> https://github.com/openstack/neutron/blob/stable/ocata/neutron/agent/linux/ip_lib.py#L982
14:46:24 <haleyb> slaweq: yes, i just know stable policy might mean security critical only
14:46:32 <slaweq> ah, ok
14:46:45 <haleyb> https://docs.openstack.org/project-team-guide/stable-branches.html
14:47:09 <amotoki> from several discussions, "critical fixes" depends on context
14:47:23 <slaweq> isn't this issue critical for LB agent?
14:47:59 <jlibosva> I would say yes if arp responder doesn't work without it
14:48:17 <haleyb> yes, and it is up to us being the maintainers, so we can override security only
14:48:18 <jlibosva> also given the complexity of the patch ...
14:48:34 <haleyb> jlibosva: whoa, those two lines are crazy!
14:48:42 <mlavalle> LOL
14:49:04 <jlibosva> two? I thought it's a oneliner
14:49:12 <haleyb> i will push the two additional changes
14:49:18 <jlibosva> oh, right, import :)
14:49:20 <haleyb> one import, one line of running code
14:49:21 <amotoki> -1/+1 means 2 lines?
14:49:33 <slaweq> amotoki: it's +2 only
14:49:51 <amotoki> another simple case :)
14:49:57 <mlavalle> let's move on... I would like to hvae 5 min at the end for aspiers
14:49:59 <haleyb> let's move on since we have :11 left
14:50:03 <jlibosva> ok, we have 10 minutes left and still some agenda :)
14:50:03 <haleyb> https://bugs.launchpad.net/bugs/1753434
14:50:04 <openstack> Launchpad bug 1753434 in neutron "Unbound ports floating ip not working with address scopes in DVR HA " [Undecided,Confirmed]
14:50:04 <aspiers> :)
14:50:21 <haleyb> i need to triage that further to see if it's still broken in master
14:50:38 <haleyb> next one
14:50:41 <haleyb> https://bugs.launchpad.net/bugs/1754695
14:50:42 <openstack> Launchpad bug 1754695 in neutron "Incorrect state of the Openflow table" [Undecided,New]
14:51:02 <haleyb> latest comment on that was that restarting neutron_openvswitch_agent container fixed issue
14:51:31 <haleyb> anyone willing to take a closer look?
14:51:39 <mlavalle> I will
14:51:44 <haleyb> sold!
14:52:16 <haleyb> last untriaged is https://bugs.launchpad.net/bugs/1754600
14:52:17 <openstack> Launchpad bug 1716043 in neutron "duplicate for #1754600 no CLI for quota details" [Low,In progress] - Assigned to Slawek Kaplonski (slaweq)
14:52:27 <haleyb> but actually see that was just marked a duplicate
14:52:40 <jlibosva> cool :)
14:52:54 <jlibosva> thanks haleyb for updates
14:53:00 <jlibosva> just a reminder that bzhao is a bug deputy for this week
14:53:09 <haleyb> mlavalle: there were two new rfes as well for drivers team
14:53:12 * mlavalle will talk to him tonight
14:53:14 <jlibosva> amotoki: boden do you have any updates to docs, cli or lib or can we skip the topic?
14:53:25 <mlavalle> haleyb: yeap, I saw them
14:53:36 <boden> jlibosva: well considering time I guess we can skip.. I can follow-up offline as needed
14:53:42 <jlibosva> boden: thanks :)
14:53:51 <jlibosva> sorry about my bad timing
14:54:04 <jlibosva> #topic On demand agenda
14:54:05 <amotoki> on SDK, i am preparing an etherpad page on gap between neutronclient and openstacksdk https://etherpad.openstack.org/p/neutron-openstacksdk-migration
14:54:22 <jlibosva> oh, sorry amotoki for changing topic already
14:54:26 <amotoki> it is WIP. I would like to double check with slaweq after i check it.
14:54:30 <mlavalle> amotoki: thanks. good initaitive :-)
14:54:35 <amotoki> jlibosva: no problem
14:54:36 <jlibosva> thanks amotoki
14:54:47 <jlibosva> aspiers put an agenda item here https://wiki.openstack.org/wiki/Network/Meetings#On_Demand_Agenda
14:54:52 <jlibosva> aspiers: stage is yours :)
14:54:56 <aspiers> thanks
14:54:59 <aspiers> hi all
14:55:03 <slaweq> amotoki: sure
14:55:17 <aspiers> details are in the link above really, but this idea came out of discussions at the PTG
14:55:29 <aspiers> I have written a Python module git-deps which can predict the complexity of porting tasks
14:55:35 <rubasov> hi
14:55:39 <aspiers> #link https://github.com/aspiers/git-deps#use-case-1-porting-between-branches
14:55:55 <aspiers> so question is, would it be useful to include this in neutron's stable branch checks pipeline
14:56:03 <aspiers> sorry, I mean in master
14:56:23 <aspiers> to get a heads up on the impact of a change in master on potential backporting work
14:56:44 <amotoki> what complexity does it check?
14:56:49 <rubasov> I'm here (together with my colleague lajoskatona) to help with the minimum bandwidth stuff this cycle
14:56:49 <aspiers> I talked to Tony Breeds about it and he suggested neutron might be a good place to try it first, since you are apparently very proactive with backporting :)
14:57:18 <aspiers> amotoki: it checks how many dependent commits are required in order for a patch to apply
14:57:22 <rubasov> we couldn't be there on the ptg unfortunately
14:57:23 <jlibosva> it would be good to hook it with launchpad tags like "queens-backport-potential"
14:57:45 <rubasov> but from other colleagues (working with nova) I know the overall plan has changed a lot
14:58:00 <amotoki> aspiers: interesting initiative
14:58:15 <aspiers> if people are interested in trying it, I'd be very happy to help
14:58:18 <rubasov> does anybody plan to re-spin the neutron-side spec for minimum bandwidth based on the current thinking?
14:58:34 <mlavalle> rubasov: let's you an I talk in the neutron channel
14:58:38 <aspiers> it's more or less ready for use right now, would just need wrapping in a zuul job I guess
14:58:50 <amotoki> it seems two topics are discussed in paralell :(
14:59:16 <mlavalle> aspiers: would it be a job in the check queue?
14:59:20 <rubasov> will take this topic offline with mlavalle, thanks
14:59:21 <aspiers> mlavalle: yes
14:59:32 <aspiers> mlavalle: well, it could be run manually too
14:59:37 <aspiers> even during development
14:59:38 <jlibosva> we can discuss details in the channel or perhaps summarizing the idea to ML about what aspiers talks about would be good?
14:59:51 <jlibosva> as we're running out of time ..
14:59:52 <aspiers> jlibosva: fine by me
14:59:58 <jlibosva> aspiers: thanks a lot!
15:00:02 <jlibosva> it sounds interesting initiative :)
15:00:02 <mlavalle> yeah, let's start a ML thread
15:00:15 <aspiers> I can hop over to neutron channel now
15:00:18 <jlibosva> rubasov: thanks for raising that topic too :)
15:00:23 <jlibosva> and thanks everybody for showing up
15:00:24 <mlavalle> yes, I founf it interesting in the PTG and asked aspiers to share with the team
15:00:26 <jlibosva> #endmeeting